|
[DPRG] Balancing robot difficulties
Subject: [DPRG] Balancing robot difficulties
From: Randy M. Dumse
rmd at newmicros.com
Date: Mon Jul 2 10:58:09 CDT 2007
Luke Steele asked: Monday, July 02, 2007 9:26 AM
>I'm hoping someone reading this will be able to help...
> So what's the problem? Although it does balance,
> it does so in quite an agitated fashion, making
> small but rapid oscillations around the balancing
> point.
To me it looks like it is balancing quite well. You are probably
closer to your goal than you think.
Your bot falls over because the top speed of the wheels is
limiting, and can not get the bottom of the bot under and ahead
of the top of the bot. This might be a tuning issue, or it might
be a weight distribution issue, or the Tamiya twin might not
have enough power.
Here's another idea. Your top speed will be limited by the
traction you have between your wheels and the surface they are
turning against. If there is slippage there, you might be trying
to accelerate to speed, and still not getting the bottom of the
bot moving faster than the top is falling. Some rubber bads
around those CD disk wheels might be a tremendous improvement.
Where are your batteries? Batteries tend to be heavy, and should
be placed at the top of the bot. Also making the bot taller,
with more mass at the top will make an easier balance. The movie
makes it difficult to tell too much, but it appears the
microprocessor (a light item) is at the top of the bot. Like it
is easier to balance a hammer with the head up (top heavy), and
it is easier to balance a broom than an pencil (length) the same
applies to this sort of bot. More mass higher will be easier to
balance.
> 1. Backlash in the gearbox/transmission.
David's encoders were mounted on the motor before the gearbox.
Since you are using a Tamiya twin, and it is very hard to get to
the motors, I'm wondering if you are encoding at the motors, or
on the output shafts. If it is the output shafts, you may be
over driving the motors. In otherwords, you are putting in
corrections on the motors, and the results don't show on the
output shafts until your PID has over responded, winding up to
during the lag as the gear backlash comes out.
This is a very common control problem, and is evidenced by way
too much motor activity in "hunt".
Are you using quadrature? You mentioned "speed" of the motors
only.
> 2. Some short-coming of my motor control. the motor
> control and balance PD loops are running at 40Hz.
40 Hz is low, but probably sufficient. What is your PWM rate?
I'm hoping at least a few thousand Hz?
> 3. Poorly calibrated PD loop:
This is a better bet than you might first think. Tuning in
control systems is very important. Oscillation is often a sign
of a D or I term turned up too high.
Randy
www.newmicros.com
More information about the DPRG mailing list
|
